On the use of a mixed binary join tree for exact inference in dynamic directed evidential networks with conditional belief functions
Résumé
Dynamic directed evidential network with conditional belief functions (DDEVN) is a framework for reasoning under uncertainty over systems evolving in time. Based on the theory of belief function, the DDEVN allows to faithfully represent various forms of uncertainty. In this paper, we propose a new algorithm for inference in DDEVNs. We especially present a computational structure, namely the mixed binary join tree, which is appropriate for the exact inference in these networks.